Exploring semantic knowledge base solutions

The WikiLibrary Manifesto

The WikiLibrary Manifesto

Our vision is to create a reliable, machine-readable, collaboratively maintained Linked Open Data network for the arts, culture and science as a solid base for FAIR knowledge. Yet fully applying the FAIR data principles to knowledge requires a shared framework. The signers of the Wikimedia Library Manifesto will work together to achieve the vision of a common knowledge graph embedded in a shared framework. They will do so by respecting the principles listed below and by taking specific measures to fulfil their vision, also outlined below.

Principles

  • Promoting free licenses for data and their software environment.

  • Shaping spaces where diverse communities thrive. (Community gardening).

  • Providing structured data based on FAIR data principles in order to be able to transparently transform data into information to create FAIR knowledge.

  • Promoting common core standards created consensually and collaboratively.

  • Providing open governance structures and embedding them into existing systems.

  • Dedicating resources to obtain user interfaces that are accessible to and userā€‘friendly for everybody who wants to contribute and actively care for data and knowledge.

  • Fostering data literacy in the digital transformation on the three stages: data, information and knowledge.

This initiative is driven from Germany. @Sebastian this seems on first glance a very interesting take on things :slight_smile:

1 Like

Yes, thank you!
I wrote yesterday https://twitter.com/sl007/status/1356160095143546882
and will add the link now.

1 Like

Please note that only institutions will appear. I co-signed but how about if
[Proposal]
the W3C Social CG signs the manifesto as institution ?

@rhiaro

Intressting, I currently discussing with https://www.inove.network and https://kartevonmorgen.org (KVM) a ā€œWiki des Wandelsā€ (Wiki of Change). There was one https://wandel.wiki but it seems to sleep.

And yesterday Alex from https://www.m4h.network agreed to set up a ā€œplaygroundā€ Semantic Media Wiki instance for us.

We want to somehow combine inove.network, KVM and https://glossardeswandels.de. There are certainly parallels to our LinkedOpenAgents (LOA) project.

Exciting time that gets even more exciting every day

But is it a good idea to have THAT ONE Wiki ? Is there something like a federated wiki :wink:

Very cool. Inove, KVM-like initatives are very interesting. Thereā€™s also the Dutch site Nudge, but they are a commercial entity.

Ultra interested to know more about it, and your plans.

This was also what popped up with me immediately when I found the resource, and what Iā€™ve discussed elsewhere too. I think it is very interesting to create a #fediversity:fediverse-futures topic about that (see Fediverse Futures idea explained). Do you think thatā€™d be interesting for you to participate in elaborating?

Hi aschrijver,
i understand what you mean by ā€œfediverse futures idea explainedā€. but right now i have absolutely no spare capacity. Iā€™m very new to linked data and federation and still struggling with many concepts. and havenā€™t even discovered a fraction yet. so today i dove a bit deeper into json-ld and came across hydra. i now also understand the background of json-ld, although iā€™m already well used to turtle and json-ld is more of a burden for me at the moment.

I am working on making organisations and events interchangeable between different platforms. Linked Open Agents (LOA). And in another project I am trying to build a minimal AP C2S. I have initially convinced the client that it is the right thing for them. Now it just has to survive the first use cases in the prototype well so I can push the added value.

I hope this works out well and we get a first ā€œminimalā€ spring-boot java AP C2S implementation based on rdf4j as OSS. So Iā€™m trying to understand AP more and more. Iā€™ve been trying to understand it for about a year now and itā€™s getting clearer every day.

And the semantic wiki is again a kind of sub-project that I want to take in the right direction. LOA and the wiki are still crying out for a way to categorise and rank things. Iā€™m in talks with kvm, glossary of change and wikirate.org about this.

So tooooooo many tasks for now :wink:

innercircles is nice, but i think i miss something ?!

The wiki thing is currently discussed here (currently in german, but i think we can try to switch to english):

I highly understand this. I am basically in the exact same situation. Also a bunch the fields of study you mention here, are on my list as well.

Donā€™t want to go off-topic too much, but this might be a PITA in type-safe Java. I started building an AP library based on Java and Vert.x (which I like a lot, more than spring-boot). You can look at Smithereen by @grishka that is written in Java. And this old ActivityStreams reference implementation is also still interesting.

Thanks for the links. I will track them with interest. I should have mentioned that the link to innercircles just points to a teaser :slight_smile: I use Matrix and Fedi, you can contact me there, or DM here for more info.

It indeed is ā€” I wrote a more-or-less complete JSON-LD processor in Java and, to put it lightly, didnā€™t enjoy it. Thereā€™s too much instanceof and Objects all over the place because the spec often calls for variables that are capable of storing arrays, JSON objects, and primitives, simultaneously. I might rewrite it to use Gson at some point in the future to make the code a bit less atrocious.

1 Like

FYI. On the Feneas forum there is this post: Project proposal: Wikipedia for Linked Data.

At Humane Tech Community my (inactive) co-admin is Samuel Klein, who is board member of The Underlay, part of MITā€™s not-for-profit Knowledge Futures Group.

Hereā€™s why The Underlay exists:

"The creation of richer collections of machine-readable knowledge is inevitable. What is not inevitable is that such knowledge will be connected in a meaningful way and be freely available as a public resource. We are at a fork in the road. The transition to machine-mediated access could consolidate our dependence on a few large commercial intermediaries, or alternatively, it could be built as open infrastructure, as a public good. "

Hereā€™s what The Underlay already (claims to) already solve:

  • how to represent complex assertions
  • how to represent the Underlay in a way that is independent of human language
  • how to indicate and verify provenance
  • how to recognize when separate sources are making assertions about the same entity
  • the technical architecture required to store and disseminate the Underlay
  • how to protect the Underlay from damage by imposters, spammers and other bad actors

A cryptpad written by Simon Grant of P2P Foundation that is worth referencing in this threadā€¦

Commons P2P wiki requirements

Two relatively recent developments caught my attention. One is Ward Cunninghamā€™s idea of the Federated Wiki; the other is the recently popular commercial software, Roam Research. I donā€™t see either as the answer, but both have interestingly different functionality.

The question I want to address here is: what functionality do we need to implement an effective ICT system that can embody a healthy, growing, live, knowledge commons ā€“ of the kind that could be used for our shared purposes? But this page is not about listing and refining those purposes, integrating all the relevant writings: that belongs elsewhere.

If this page could start to develop into a list of requirements that has broad agreement, then we could move on in two ways:

  1. we could invite people to build a reference implementation, along with clarifying the relevant shared protocols, APIs and standards;

  2. we could invite developers of existing open systems to work on incorporating this functionality into their systems.

A related topic on the Solid community forum:

@flancian is building the knowledge base Agora, discussed in Show HN: The Agora ā€“ an experimental social network | Hacker News (first HN comment gives a summary of Agora) and has very interesting concepts (though I didnā€™t deeply dig into them yet, and it can be a bit hard / overwhelming to navigate). Btw, Flancian is exploring a connection to [[underlay]] :


@VictorVenema is outlining a use case related to knowledge collection in Bulding information sources with lasting value.

Presenting yesterday at the NGI Linked Data webinar was AtomGraph with their: LinkedDataHub - The Knowledge Graph management system

LinkedDataHub is open source software you can use to manage data, create visualizations and build apps on RDF Knowledge Graphs

Hereā€™s another FedWiki page that explains more about how it works, and what the relationship is to ā€˜commoningā€™ by Silke Helfrich

Federated Wiki

A single-page browser application reads from many sites at once and saves changes in that browser (see foto).
Users can host their own sites, where they edit and organize their knowledge. At the same time they can share back to the federation.
The tools for refactoring and sharing are "hard to learn but easy to use. ā€œDrags 'n dropsā€ and ā€œforksā€ are omnipresent. The software, published as a Node.js package, runs on a variety of platforms. It can be used on a personal laptop or an industrial server.

[ā€¦]

How do they work?

Ward Cunningham, the ā€œfatherā€ of the fed/wiki describes:

  • Federated Wiki ā€” Federated Wiki sites share pages circulating within a creative commons.

FedWiki is the next wiki-generation. It is a software that allows everybody to create individual wiki pages while circulating its content in a global - federated - environment. It opens multiple paths for individual knowledge organization while demanding equal responsibility in creating a knowledge commons . This knowledge commons emerges out of thousands of wiki sites people freely contribute to.

FedWiki invites everybody - called ā€œvisitorsā€ in FedWiki language - to go around the wiki-pluriverse and take any point of view, without having to conform with the version presented through - say - a Wikipedia page.

Which Core Dimensions of Commoning are fostered?


@bashrc started an interesting discussion on AP-federated wikiā€™s in which a bunch of notable projects were mentioned:

On the same discussion thread Mike Hales in response to @bhaugen posted the following (CC BY-SA-4.0):

I agree wikis are important and federation across wikis too. Sorry, Iā€™m not able to easily imagine what the characteristics of a system of wikis might be, built under an #activitypub protocol. That would be quite a thought experiment - worth doing properly Iā€™d say. But I can try to outline the important characteristics of federated wiki imho. Aka ā€˜wikiā€™.

In contrast with wiki(pedia), wiki is for individual rather than collective authorship/ownership. The pages or paragraphs of a wiki site can be forked without limit within the wiki community. Ward Cunningham (wiki creator/guru) calls this ā€˜a chorus of voicesā€™ in contrast with the wiki-for-consensus that most people are familiar with.

wiki has affordances that enable a community to form around (using) the membersā€™ own wikis, and publish work to each other with an intention to collectively evolve an understanding or a body of content, in wiki. This is sometimes called ā€˜poddingā€™. A parallel (sync or async) communication channel is needed too - messaging, email, videoconferencing, etc.

wiki has affordances for observing what other known writers have been at work on (signified by page edits) and for constructing a ā€˜neighborhoodā€™ of sites containing resources believed to be useful in the context of the present wiki. Within the neighborhood, free-text search is available. Any found page (or paragraph within it) can be forked to the present wiki. Thus wiki is ā€˜an awareness engineā€™.

wiki is profoundly hypertextual - not only links across and within sites but also endless forking of content and versioning of pages. A page maintains a complete version history, including sources of content. Hyperspatiality makes ā€˜wiki spaceā€™ hard for many folks to navigate in. A feature of wiki UI is narrow pages laid side-by-side in the window. This helps in flattening hyperspace. But this is also confusing for many folks used to multiple browser tabs

wiki comes out of the Smalltalk community. Its original programmer Ward Cunningham also programmd the first wikipedia-type wiki (called ā€˜wiki wikiā€™). wiki is an attempt by Ward, 30 years later, to provide tools for creative authorship and active making, rather than librarianship and the authority of professional fields. In this regard it may be closer to Tiddlywiki for example (eg used for personal notebooks) than to wikipedia.

Coming from Smalltalk (and, say, Hypercard), and from Ward as an ace creative programmer/problem-solver, wiki includes affordances for computation within the page, across pages in a wiki site, across sites in a farm (a server) and across servers in the wikiverse. wiki is absolutely not just a text engine, itā€™s a federal computing environment.

Significant 2021 innovations from wikiā€™s geeky core community include plugins that do computation from data on a siteā€™s pages; also drawing visual maps of relationships within a web of linked pages (and other graph applications); also organising data harvested in large-scale geographical spaces. The tooling capability of wiki is enormous. Again this makes it harder for ā€˜simpleā€™ users to really mobilise the power. But to ā€˜get itā€™, readers must be writers.

TBH the wiki community is pretty geeky, largely comprising sysadmins who self-host farms. Experiments are only just beginning, in creating hosted wiki farms for non-geeks to use in a routine, ā€˜vernacularā€™, SaaS kind of way - as personal notebooks say, or as vehicles for explicit federating across a community of collaborating co-makers (eg community organisers, educators, coop development folks, activist networks, designers).

FYI my wikis can be found from this page of ā€˜rostersā€™: Rosters

Another listing is here: What goes on

Some are work-in-progress. Some are scrapbooks. Some are resource catalogues to support projects. Some are drafts of books. Some are thought spaces for reflecting on issues (components of a digital toolstack for organisers, the institutional labour of activist education aka formaciĆ²n, ā€˜making the living economyā€™).

My own collection of howtowiki links is here: wiki literacy

I surely have missed some features, might tweak this series if a conversation develops.

I feel wiki is a very serious collaborative text-based environment; also a serious distributed computing environment. Intrinsic complexity makes it puzzling for many folks. This kind of issue is only just being engaged in the wiki core community, with a view to making wiki a space of vernacular tools for conviviality.